Available online at www.sciencedirect.

com

ScienceDirect
Procedia Technology 24 (2016) 70 – 76

International Conference on Emerging Trends in Engineering, Science and Technology (ICETEST

- 2015)

Simulation of Forward Osmosis Using CFD

Shyama Pankaja, N. Sajikumarb *, Roopesh Kaimalc
a
Graduate student Dept. of Civil Engineering, Govt. Engineering College,Thrissur, India
b
Professor,Dept. of Civil Engineering, Govt. Engineering College,Thrissur, India
c
Graduate student Dept. of Mechanical Engineering, Govt. Engineering College,Thrissur, India

Abstract

If fresh water and salt water are separated by a membrane, the fresh water would flow across the membrane from fresh water
chamber to salt water chamber in a forward osmosis process. Forward osmosis and reverse osmosis processes have been used for
many new applications like wastewater treatment, food processing industry, and desalination of seawater. Computational Fluid
Dynamics (CFD) software, viz., FLUENT software is used for simulating FO process. The water moved across the membrane
from fresh water side to salt water chamber indicating that FO was well represented by FLUENT.
© 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
© 2016 The Authors.Published by Elsevier Ltd.
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-reviewunder
Peer-review underresponsibility
responsibility of the
of the organizing
organizing committee
committee of ICETEST
of ICETEST – 2015.
– 2015

Keywords: Forward Osmosis (FO); Computational Fluid Dynamics (CFD); FLUENT.

1. Introduction

Modelling of osmosis process, both experimental and numerical has become a hot subject of research over
the last few decades. Osmosis is defined as the phenomenon of water flow through a semi-permeable membrane
which does not permit the transport of salts or other solutes across it. More specifically, it is the movement of water
through a semi-permeable membrane from one side of the membrane with low solute concentration to the other side
of the membrane with high solute concentration. Osmosis is a reversible thermodynamic process (Cath et al. 2006;
Sabah et al., 2013). That is, by proper adjustment of external pressure on the solution side, the direction of water
flow across the membrane can be controlled. There are three types of osmotic processes: Forward Osmosis (FO),

* Corresponding author. Tel.: +0-000-000-0000 ; fax: +0-000-000-0000 .

E-mail address: [email protected]

2212-0173 © 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the organizing committee of ICETEST – 2015
doi:10.1016/j.protcy.2016.05.011
 Shyama Pankaj et al. / Procedia Technology 24 (2016) 70 – 76 71

Reverse Osmosis (RO) and Pressure Retarded Osmosis (PRO) (Gruber et al., 2011). In FO, the driving force for
transport of water through the membrane is the osmotic pressure difference (Δπ) across the membrane.. Flux in FO
is from low concentrated salt solution to high concentrated salt solution when applied pressure difference is zero, i.e.
ΔP=0. In RO, the driving force for the transport of water through the membrane is the applied pressure. Flux in RO
is in the opposite direction of FO, i.e. from highly concentrated salt solution to low concentrated salt solution. In
PRO, the hydraulic pressure is applied to the draw solution but is less than the osmotic pressure existing between the
solutions in the system; as a result, the water flux is in the direction of the concentrated draw solution similar to
Forward Osmosis. The main advantage of FO over RO and PRO is that it is more cost effective due to the lack of
need for hydraulic pressure (Cath et al., 2006).
Modelling of complex flow patterns in membrane systems using Computational Fluid Dynamics (CFD)
have become increasingly popular as it provides more robust approach capable of including many flow parameters
(Cath et al. , 2006; Gruber et al. 2011). Studies on modelling of Osmosis using CFD have utilized User Defined
Function (UDF) facility of FLUENT or source term adjustment for modelling the process. In this study, the process
is represented by the cell zone definition. In order to substantiate this proposition, CFD simulation is carried out
without using any of the methodologies mentioned earlier.
The objective of this paper is to model Forward Osmosis system with an asymmetric membrane using CFD
model. For this ANSYS FLUENT 14.5, CFD based software have been used. The membrane module consists of a
thin non-porous skin dense layer called skin, on top of a porous support layer. The porous layer supports the
membrane but they reduce the efficiency of the membrane. Multiphase mixture model was used.

1.1. Theory

The general equation describing water flux in an osmotic process is given by

Jw = A∙(Δπ − ΔP) (1)

where Jw is the water flux through membrane, A is the water permeability coefficient of the membrane, Δπ is the
osmotic pressure difference and ΔP the hydraulic pressure differential across the membrane (Fluent 2006; Gruber et
al. , 2011).

Osmotic pressure (ΠOSMOTIC) can be calculated by the Van’t Hoff equation:

Π = 2∙CNaCl ∙R∙T (2)

where R is the gas constant, T is the absolute temperature and CNaCl is the concentration of NaCl (Gruber et al. ,
2011).

All CFD software solves seven equations, these are: Equation of conservation of mass, three momentum
equations in three directions, Energy equation, Equation of state and Equation of enthalpy.

The Equation of state can be expressed as:

P = ρ*R*T (3)

where P is the pressure, R is the gas constant,T is the absolute temperature and ρ is the density.
For a mixture, density becomes density of the mixture ρmix. ρa for component A in mixture is expressed as
the ratio of mass of component A by total volume of mixture. This represents the concentration c. The equation (3)
can be changed as,

P = c *R*T (4)

This equation is similar to Van’t Hoffs equation for osmotic pressure π and is given by
72 Shyama Pankaj et al. / Procedia Technology 24 (2016) 70 – 76

Π= cRT (5)

Hence, the osmotic pressure can be represented by the modelling processes in FLUENT. Earlier studies on
FO using CFD have utilized User Defined Function (UDF) facility of FLUENT or source term adjustment for
modelling the process of FO. In this study, the process is represented by the cell zone definition. In order to
substantiate this proposition, CFD simulation is carried out without using any of the methodologies mentioned
earlier.

1.2. CFD Simulation

ANSYS Fluent is a CFD solver having comprehensive modelling capabilities for a wide range of
compressible and incompressible, laminar and turbulent fluid flow problems. Any type of flow can be simulated
using this solution environment for fluid problems. In order to simulate the FO, two chambers separated by
selectively permaeable memberane (semi-permeable) have to be created so that the brine solution and freshwater
can be kept in these chambers.
The model consists of a fresh water chamber (75mm x 25mm); salt water chamber (75mm x 25mm)
separated by a semi-permeable membrane (2.5mm x 25mm). The basic geometry is shown schematically in Fig.1.
Flow direction of brine and fresh water are indicated in the figure. In addition a flow will be established across the
membrane (i.e. perpendicular to flow direction) if the osmosis process is developed. The semi-permeable membrane
is treated as porous zone. The geometry dimension and sea water concentration are selected from bench-scale
experiment carried out by Achilli et.al. (2009).

Fig. 1 Created geometry for the analysis.

The geometry was discretised into small control volumes by using ANSYS Mesh Modeller or module. The
meshed geometry had about 187200 elements and 200718 nodes.
 Shyama Pankaj et al. / Procedia Technology 24 (2016) 70 – 76 73

Fig.2 Meshed geometry

For modelling the flow, pressure based solver with transient time was chosen. The selcted models were
viscous laminar model and multiphase mixture model. The mixture model is a simplified multiphase model that can
be used to model multiphase flows where the phases move at different velocities.

The continuity equation for the mixture is expressed as,

డ
ሺߩ݉ሻ ൅ ‫׏‬Ǥ ሺߩ௠ Ǥ ‫ݒ‬Ԧ௠ ሻ ൌ ݉ሶ (6)
డ௧

Where‫ݒ‬Ԧ௠ is the mass-averaged velocity:

σ೙ ሬԦೖ
ೖసభ ఈೖ ఘೖ ௩
‫ݒ‬Ԧ௠ ൌ ఘ೘
(7)
Where ߩ௠ is the mass-averaged density:
௡
ρ୫ ൌ ෍ α୩ ρ୩ (8)
௞ୀଵ

Where Ƚ୩ is the volume fraction of phase k; m represents mass transfer due to cavitation or user-defined mass
sources.
The momentum equation for the mixture can be obtained by summing the individual momentum equations for all
the phases. It can be expressed as:

డ ் ሻሿ
ሺߩ௠ ‫ݒ‬Ԧ௠ ሻ ൅  ‫׏‬Ǥ ሺߩ௠ ‫ݒ‬Ԧ௠ ‫ݒ‬Ԧ௠ ሻ ൌ െ‫ ݌׏‬൅  ‫׏‬Ǥ ሾߤ௠ ሺ‫ݒ׏‬Ԧ௠ ൅  ‫ݒ׏‬Ԧ௠ ൅  ߩ௠Ǥ ݃Ԧ ൅ ‫ܨ‬Ԧ ൅  ‫׏‬Ǥ ሺσ௡௞ୀଵ ߙ௞ ߩ௞ ‫ݒ‬Ԧௗ௥ǡ௞ ‫ݒ‬Ԧௗ௥ǡ௞ ሻ (9)
డ௧

where n is the number of phases,‫ܨ‬Ԧ is a body force, and μm is the viscosity of the mixture and is given by:
ߤ௠ ൌ σ௡௞ୀଵ ߙ௞ ߤ௞ (10)
‫ݒ‬Ԧௗ௥ǡ௞ is the drift velocity for secondary phase k given as:
‫ݒ‬Ԧௗ௥ǡ௞ ൌ ‫ݒ‬Ԧ௞ െ ‫ݒ‬Ԧ௠ (11)
The energy equation for the mixture takes the following form:
డ
σ௡௞ୀଵሺߙ௞ ߩ௞ ‫ܧ‬௞ ሻ ൅  ‫׏‬Ǥ σ௡௞ୀଵሺߙ௞ ‫ݒ‬Ԧ௞ ሺߩ௞ ‫ܧ‬௞ ൅ ‫݌‬ሻሻ ൌ ‫׏‬Ǥ ሺ݇ୣ୤୤ ‫ܶ׏‬ሻ ൅ ܵா (12)
డ௧

Where keff is the effective conductivity.

74 Shyama Pankaj et al. / Procedia Technology 24 (2016) 70 – 76

The first term on the right-hand side of represents energy transfer due to conduction.SE includes any other
volumetric heat sources.
௣ ௩ೖమ
For a compressible phase, ‫ܧ‬௄ ൌ ݄௞ െ ൅ (13)
ఘೖ ଶ

For an incompressible phase, ‫ܧ‬௞ ൌ ݄௞  (14)

Where hk is the sensible enthalpy for phase k.

From the continuity equation for secondary phase p, the volume fraction equation for secondary phase p can be
obtained:
డ
൫ߙ௣ ߩ௣ ൯ ൅  ‫׏‬Ǥ ൫ߙ௣ ߩ௣ ‫ݒ‬Ԧ௠ ൯ ൌ െ‫׏‬Ǥ ሺߙ௣ ߩ௣ ‫ݒ‬Ԧௗ௥ǡ௣ ሻ (15)
డ௧

The semipermeable membrane was treated as porous zone and boundary condition was chosen as interior.
An asymmetric semi-permeable membrane of cellulose acetate with a porous support layer of 2mm and skin of
0.5mm thickness was used for the simulation.
Forward Osmosis trial was modelled by taking the porosity of membrane as 0.05%, the NaCl concentration
as 35g/l. The average steady flux through the membrane was obtained as 5.8 x 10 -5 m/sec. The flow took place from
fresh water chamber to brine water chambers though the hydraulic pressure (without considering osmotic pressure)
was slightly higher on the brine solution side, owing to higher density. Since the flow direction is from the fresh
water to brine solution against the hydraulic pressure (without considering osmotic pressure) it can be inferred that
the osmotic process is being properly represented. If the osmotic process was not represented, the flow would have
taken place in the opposite direction. In order to substantiate, the salt concentration in the both sides just before the
starting of computation and at different stages are shown in Fig.2. It may be noted that the salt concentration on the
fresh water side remains to be constant while on the other side, it diminishes owing to the water movement from
fresh water chamber to salt water chamber. It is seen that the average salt volume fraction reduces from 0.035 to
0.0172 and approaches a constant value. This indicates the Forward Osmosis is being well represented. Average
concentrations with respect to time on the brine solution side are given in graph (Fig.3).

(a)
 Shyama Pankaj et al. / Procedia Technology 24 (2016) 70 – 76 75

(b)

(c)

Fig. 3 Variation salt volume fraction in both the chamber (a) at 0 Sec, (b) after 5Sec, (c) after 30 Sec.
76 Shyama Pankaj et al. / Procedia Technology 24 (2016) 70 – 76

Fig. 4 Average NaCl Volume Fraction Vs Time in brine water chamber.

Here, the variation of average volume fraction within the brine water chamber has decreased from the
initial value of 0.035g/l to a constant 0.0172g/l, indicating that a steady state has attained. Hence, a constant flow of
fresh water through the semi-permeable membrane takes place. Thus it can be concluded that the osmotic process is
well represented by the developed mixture based viscous model.

2. SUMMARY
Process of Osmosis was simulated using CFD based software ANSYS FLUENT 14.5. An asymmetric
semi-permeable membrane of cellulose acetate with a porous support layer of 2mm and skin of 0.5mm thickness
was used for the simulation. Water flux through the semi-permeable membrane and the volume fraction of NaCl in
fresh water chamber were calculated with the help of CFD Post Processor Module.

3. CONCLUSIONS
The following conclusions are derived from the study:
x The process of Forward Osmosis was well modelled using CFD software FLUENT. This is indicated by
the reduction in volume fraction of NaCl in brine water side.
x The volume fraction of Nacl in the sea water chamber was found to reducing with each time step, which
indicates the penetration of fresh water through the semi-permeable membrane, though hydraulic pressure
was slightly higher on the brine solution side, owing to higher density, which is clear indication showing
that the osmotic process is taking place.

References

Achilli A., Cath Y.T., Childress E.A. (2009): ‘Power generation with pressure retarded osmosis: An experimental and theoretical investigation’,
Journal of Membrane Science, vol.343, pp.42-52.

Cath Y.T, Childress E.A, Elimelech M. (2006): ‘Forward osmosis: Principles, applications, and recent developments’, Journal of Membrane
Science.

Chou S., Wang R., Shi L., She Q, Tang C., Fane G.A. (2012):’ Thin-film composite hollow fiber membranes for pressure retarded osmosis (PRO)
process with high power density’, Journal of Membrane Science, Vol.389, pp.25-33.

Fluent, 2006. Fluent User’s Guide, ANSYS Inc, Lebanon.

Gruber M.F., Johnson C.J., Tang C.Y., Jensen M.H., Yde L., Hélix-Nielsen C. (2011): ‘Computational fluid dynamics simulations of flow and
concentration polarization in forward osmosis membrane systems’, Journal of Membrane Science, Vol. 379 , pp. 488– 495.

Sabah M., Atwan F.A., Mahood B.H, Sharif A. (2013): ‘Power generation based on Pressure Retarded Osmosis: A design and an optimisation
study’, International Journal of Application or Innovation in Engineering & Management (IJAIEM), Vol. 2, pp.68-74